MIT Scales Up Homework Reviews

MIT researchers have developed a new system to help instructors review computer science homework assignments in courses with hundreds or thousands of students.

While instructors can easily create automated tests to determine whether students' computer programs produce the correct results, it's challenging for instructors to conduct a detailed analysis of the programming methods the students used when a course has hundreds of students, or in the case of some online courses, thousands of students.

The new system, called OverCode, analyses and compares students' computer code, groups solutions that use the same techniques and creates a program template for each technique. The system then displays the templates side-by-side and highlights any differences between them. Instructors can view a list of student programs that use each template.

The MIT researchers evaluated OverCode in two usability studies and will present the results at the Association for Computing Machinery's Conference on Human Factors in Computing Systems in April. The studies involved 24 experienced programmers reviewing three different programming assignments from thousands of students using both OverCode and another tool that displays solutions one at a time. "When assessing the simplest of the three assignments, the subjects analyzing raw code performed as well those using OverCode," according to information from MIT. "For the most difficult of the three assignments, however, the OverCode users covered about 45 percent of student responses, while the subjects analyzing raw data covered only about 9 percent."

According to Elena Glassman, an MIT graduate student in computer science and engineering and first author on the paper, the results of the study suggest that OverCode is particularly useful for analysing more complex computer programming assignments. The researchers also said they think the system could help instructors of large online courses provide generalized feedback to more students and use the information to improve both online and on-campus courses.

About the Author

Leila Meyer is a technology writer based in British Columbia. She can be reached at [email protected].

Featured

  • silhouette of business person facing wall of data

    Why AI Strategy Belongs in the President's Office

    Institutions that are succeeding with AI share one thing in common, and it is not a better committee, a larger budget, or a more sophisticated technology stack. It is a president who never handed off the steering wheel.

  • cloud icon with internal and external connections

    New Agentic AI Tool Analyzes Oracle Fusion and Workday Releases

    AI-powered automation platform Opkey has announced Release Advisor, a new agentic AI product aimed at helping Oracle Fusion and Workday customers analyze release updates, determine impact, and generate testing plans for their environments.

  • digital brain with network connections

    Microsoft Moving to Internally Developed AI Models in Office Apps

    Microsoft is reportedly using its own in-house artificial intelligence models to handle some workloads in Excel and Outlook, offering new evidence that the company is moving its AI strategy beyond model development and into large-scale cost reduction.

  • abstract glowing cube outlines

    Microsoft Positions Windows as an Operating Environment for AI Agents

    The recent Microsoft Build 2026 developer conference highlighted a significant shift in the company's Windows strategy. Rather than presenting artificial intelligence as a collection of standalone features, Microsoft is increasingly positioning Windows as a platform for AI agents.